Anthropic Develops Tailored AI Models for US National Security Clients

Anthropic has launched a specialized suite of AI models designed specifically for U.S. national security clients.

According to a blog post, the custom "Claude Gov" models were developed in direct response to government feedback to meet practical operational demands. Unlike Anthropic's consumer and enterprise offerings, these new models are tailored for government functions such as strategic planning, mission support, and intelligence analysis.

Anthropic notes that the models are already being used by top-tier U.S. national security agencies, with access restricted to personnel operating in classified settings. The company confirms that the models underwent the same comprehensive safety evaluations applied to all Claude AI systems.

This development aligns with Anthropic’s strategy to secure stable revenue streams through U.S. government partnerships. In November, the company collaborated with Palantir and Amazon Web Services—the cloud division of its key partner and investor, Amazon—to market its AI solutions to defense organizations.

The custom Claude Gov models demonstrate improved handling of classified content, showing reduced refusal rates when processing sensitive information. They also exhibit a deeper contextual understanding of intelligence and defense documents. Anthropic highlights the models' enhanced capabilities in national security-relevant languages and dialects, along with advanced interpretation of complex cybersecurity data for analytical purposes.

Anthropic joins several leading AI laboratories pursuing defense sector opportunities.

OpenAI is working to strengthen its ties with the U.S. Department of Defense, while Meta recently announced expanded availability of its Llama models for defense applications. Google is developing a classified-environment version of its Gemini AI platform. Separately, TechCrunch reported in early December that enterprise AI specialist Cohere is partnering with Palantir to implement its AI models.

Register Now for TechCrunch Sessions: AI with Early Bird Rates Through June 4

Secure $300 off your TC Sessions: AI pass—plus 50% off a second ticket. Join industry pioneers from OpenAI, Anthropic, and Khosla Ventures for a day of expert presentations, interactive workshops, and strategic networking. These discounted rates expire when the event begins on June 5.

Exhibit at TechCrunch Sessions: AI

Connect with 1,200+ decision-makers at TC Sessions: AI and showcase your innovations without major exhibition costs. This limited table availability ends May 9 or when spaces are filled.